@flowingboy/dsh-local-perf
持久化的 DeepSeek Harness 套件：将本地模型性能调优作为可重新安装的插件层提供（压缩、工具结果裁剪、时间上下文、云端标题路由、文本工具调用保护）——在 dsh 更新后仍然保留
插件会安装到这里；不确定时保持 web。
npx -y @deepseek-ai/dsh plugin --profile web add github:flowingboy/dsh-local-perf#b306e53ab0df4e4a567e61f44ed950b90deb98adLocal Perf 以 @flowingboy/dsh-local-perf 发布，当前版本为 0.2.0。Plugin Hub 会校验它的 manifest，并保存精确安装来源，便于复现安装结果。
Durable DeepSeek Harness plugin bundle carrying the local-model performance tuning process as a re-installable layer — so it survives dsh updates instead of living in hand-edited patch files that a version bump can wipe or whose rationale dies with the author.
Install once, re-apply forever:
# from the GitHub repository
git clone https://github.com/flowingboy/dsh-local-perf.git /tmp/dsh-local-perf
dsh plugin --profile web add file:/tmp/dsh-local-perf
# or from a checkout: dsh plugin --profile web add file:./dsh-local-perf
The bundle joins the profile's layer stack after dsh-base / dsh-web-app,
re-applies all tuning rows on every boot, and vendors its own copy of the
text-toolcall-guard plugin (self-contained, no dsh checkout required).
Everything below was learned the hard way on an M5 Max running four local OpenAI-compatible servers (Ollama / MLX / Rapid-MLX / MLX-DSpark) against the DSH web GUI. The bundle encodes the conclusions; this README preserves the reasoning.
| Date | Symptom | Root cause | Fix |
|---|---|---|---|
| 2026-08-18 | "fan spin / no response" | A 114,650-token prefill on Qwen3.5-122B — session context had grown unbounded because the web bundle disables auto-compaction | Re-enable compaction-basic at thresholdRatio: 0.6 |
| 2026-08-19 | session-title starves the interactive step | The title LLM request fires in the same second as a turn's first step; on the single-slot mlx-dspark server it queues behind the interactive generation and one of them starves past the idle budget | Route session-title-llm to the cloud model (deepseek-official / deepseek-v4-flash) |
| 2026-08-19 | mlxdspark timeouts with tools present | The mlx-dspark server buffers the whole generation and emits no data events until it finishes; its 15s SSE keepalive comments are discarded by the OpenAI SDK parser and never reset the idle watchdog | Raise timeoutMs / streamIdleTimeoutMs to 600000 |
| 2026-08-20 | tool calls appear as literal text | Local Qwen3.8-27B-8bit fell out of the structured tool_calls protocol under long tool-heavy steps and wrote <tool_call> prose the harness never executes | Ship the text-toolcall-guard plugin (vendored here) |
| recurring | hallucinated "today" | No clock context in the prompt | Enable time-context (Asia/Shanghai, 10 min refresh) |
settings.example.yaml → ~/.dsh/settings.yaml)Machine-specific (paths, ports, model ids), so the bundle carries them as a
template, not a runtime patch. Copy the llm-pi-ai section into
~/.dsh/settings.yaml on a fresh machine.
The recurring principles:
timeoutMs +
streamIdleTimeoutMs ≥ worst-case prefill + reasoning + decode. Gemma 4 31B
prefills at ~180 tok/s (system prompt + tool schemas ≈ 13k tokens → ~70s),
mlx-dspark buffers whole generations, so both budgets sit at 300–600s.cordis.patch.yml)| Row | What | Why |
|---|---|---|
time-context | per-step clock | kills hallucinated dates |
compaction-basic | auto-compact at 60% | bounds prefill; the 08-18 incident fix |
tool-result-pruner | drop stale tool results | keeps them off later requests |
command-compact | manual /compact | escape hatch |
session-title-llm | title via cloud model | keeps the local slot free for the interactive step |
plugins/text-toolcall-guard)Vendored from @deepseek-ai/dsh-text-toolcall-guard (built lib/ + src/).
When a step closes with no native tool calls but the assistant text carries
<tool_call> / <function=…> markers at line starts, the guard steers a
corrective message so the model re-issues the call natively. Bounded to
maxCorrections (2) per turn per agent; marker detection requires line-start
placement so prose that merely quotes the format is not corrected. Peers are
resolved at runtime from the dsh installation's profiles/node_modules
fallback (the designed out-of-tree-plugin path), so no registry fetch needed.
# from this directory
dsh plugin --profile web add file:$(pwd)
# verify the layer joined the stack
dsh --profile web --dump-config | grep -A3 "dsh-local-perf"
The web profile's cordis.patch.yml should then only hold rows this bundle
does NOT own (currently: none — everything moved into the bundle).
git pull # or edit locally
cd plugins/text-toolcall-guard # rebuild the vendored guard if its src changed
pnpm exec tsc -p tsconfig.json --outDir lib --declarationDir lib/types
# reinstall the layer
dsh plugin --profile web add file:$(pwd)
Published at https://github.com/flowingboy/dsh-local-perf (public, main,
topics: dsh-plugin). To re-publish after local edits:
git add -A && git commit -m "dsh-local-perf: ..."
git push origin main
# topics (once)
gh repo edit dsh-local-perf --add-topic dsh-plugin
cordis.patch.yml the perf layer (all tuning rows)
settings.example.yaml model-config template (copy to ~/.dsh/settings.yaml)
plugins/text-toolcall-guard/ vendored guard plugin (lib + src)
README.md this document — the preserved process